Why are listeners sometimes (but not always) egocentric? Making inferences about using others’ perspective in referential communication

Abstract

Theory of Mind (ToM) is the ability to understand others’ mental states, and that these mental states can differ from our own. Although healthy adults have little trouble passing conceptual tests of ToM (e.g., the false belief task [ 1 ]), they do not always succeed in using ToM [ 2 , 3 ]. In order to be successful in referential communication, listeners need to correctly infer the way in which a speaker’s perspective constrains reference and inhibit their own perspective accordingly. However, listeners may require prompts to take these effortful inferential steps. The current study investigated the possibility of embedding prompts in the instructions for listeners to make inference about using a speaker’s perspective. Experiment 1 showed that provision of a clear introductory example of the full chain of inferences resulted in large improvement in performance. Residual egocentric errors suggested that the improvement was not simply due to superior comprehension of the instructions. Experiment 2 further dissociated the effect by placing selective emphasis on making inference about inhibiting listeners’ own perspective versus using the speaker’s perspective. Results showed that only the latter had a significant effect on successful performance. The current findings clearly demonstrated that listeners do not readily make inferences about using speakers’ perspectives, but can do so when prompted.

Are you thinking what I’m thinking? Perspective-taking in a language game

“…We therefore expanded the trial-by-trial feedback in Experiment 2 so that participants were also told after unsuccessful trials what the optimal clue word would have been, and what guess people usually provide in response to the clue the director actually gave (see Fig 7 for an example of the feedback screen in Experiment 2). Given previous literature showing that explicitly instructing participants to take perspective substantially improves performance [46], the more extensive feedback should also be helpful to our participants. This is similarly supported by beneficial effects of constraining the search space (e.g., [29]), effectively showing participants what good clues are like.…”

Rethinking egocentric bias: A computer mouse-tracking study of adult belief processing.

“…Although adults are capable of inferring that another agent has a perspective or a belief that is different from their own, they often fail to make use of this ability in ongoing tasks, such as when inferring the referent of another's statement (e.g., in the widely used director task: Apperly et al, 2010;Keysar et al, 2003;Lin et al, 2010; J. J. Wang et al, 2020). Such behavior suggests a type of egocentrism: Adults may simply fail to appreciate the relevance of another person's mental state to the task at hand.…”

Action Prediction: a Multimodal Techniques Investigation of the Functional Relationship Between Belief-Tracking and Motor Processes

“…Building from Experiment 1A's methodology and procedure, Experiment understanding of what is required in the task has been proven to be a suitable motivator in getting adults to take others' perspective in cognitively effortful scenarios (Wang, Ciranova, Woods, & Apperly, 2020). Second, I introduced a reward (i.e., a candy) that could be obtained when performing an helping behaviour that was in accordance with the instructions.…”